David R. Liu

Scholar: 2000

Awarded Institution
Harvard University and Harvard School of Public Health
Department of Chemistry and Chemical Biology


Research Interests

Development of Amplifiable and Evolvable Unnatural Molecules

Combinatorial approaches towards the isolation of molecules with new function have been employed with great success over the past decade. These efforts have lead to the isolation of nucleic acids, proteins, and small molecules with novel binding, catalytic, or medicinal properties out of libraries of hundreds to trillions of variants.[1] Nucleic acids and proteins, unlike their small molecule counterparts, share a critical feature which explains their dominance among solutions to complex chemical problems: these molecules can be amplified. This feature allows, in theory, a single molecule out of billions to be isolated by iterated selection and amplification, providing exquisite sensitivity. Equally important, amplification is a prerequisite for molecular evolution, and researchers have demonstrated extensively that nucleic acids and proteins which initially lack or only weakly possess desired activities can be mutated, amplified, and reselected to afford molecules with greatly enhanced properties.[1] Yet nucleic acids and proteins, while amplifiable, lack the flexibility which chemists routinely exert on synthetic small molecules and thus are restricted to a limited set of chemical and biological properties. We are interested in designing amplifiable unnatural polymers, thus combining the major benefits of natural molecules with the design flexibility of synthetic molecules. DNA-templated synthesis provides a promising starting point towards this goal.[2-9] Libraries of amplifiable unnatural polymers would be suitable for sequencing, solid-phase or solution-phase selections, random or directed mutagenesis, and homologous recombination using DNA shuffling[1,10,11] and therefore can be evolved over iterated rounds of selection and amplification using the same extremely powerful genetic methods employed by Nature during the evolution of life. This strategy may lead to the isolation of entirely new classes of unnatural molecules which selectively bind molecules or which catalyze chemical reactions. Characterization of these molecules will provide important insight into the ability of unnatural polymers such as polycarbamates, polyureas, polyesters, polysaccharides, or peptide nucleic acids bearing novel side chains to form higher-order structures with efficient binding or catalytic properties. We are also applying this approach to the generation of amplifiable non-polymeric molecules through the development of chemistry for general DNA-templated bond formation on solid support as well as in solution. The development of such systems may allow the direct evolution of new generations of small molecule ligands and drugs, an approach which may prove more effective than traditional cycles of compound screening and analog synthesis.

Molecular Evolution of Proteins

We are also interested in the evolution of natural molecules (proteins and nucleic acids) using in vivo selections by coupling the survival of a cell with the solution to a chemical problem of interest. Initial efforts will focus on (i) the evolution of nucleases with arbitrary single site per genome cleavage specificity, (ii) the evolution of new trans intein pairs with exclusive pairwise protein splicing activity, (iii) the evolution of artificial allosteric proteins which are activated or inactivated by arbitrarily chosen small molecules, (iv) the evolution of recombinases with new substrate specificities, (v) the evolution of fibril-forming and fibril-resistant prions and (vi) the design and evolution of artificial transcription factors with tailor made specificities. In each case, the evolution of proteins with desired properties will both answer basic science questions in protein-DNA, protein-protein, or protein-small molecule recognition as well as provide powerful new research tools. Evolved nucleases, for example, may lead to virus-resistant cell lines and enzymes capable of selective allelic destruction in vivo. Inteinevolution may provide a general method of intracellular protein recombination to create libraries of more than 10^(16) protein variants in vivo. Artificial allosteric proteins may serve as chemical sensors for the detection of small molecules inside or outside of living cells. Engineered and evolved transcription factors may be used to activate or repress the expression of any gene in vivo or in vitro. We are also interested in defining the scope of biopolymeric catalysis and in expanding our ability to generate complex protein- and RNA-based catalysts by the stepwise evolution of new protein catalysts from catalytic RNAs and of new ribozymes from existing protein enzymes. We are currently developing both an in vitro and an in vivo method for selecting catalytically active protein-RNA complexes.

Biologically Inspired Synthetic Methodologies

A third area of focus lies in the development of synthetic methodologies which employ new approaches, often inspired by Nature, to address existing problems in interdisciplinary areas spanning chemistry and biology. For example, we are developing a new "protecton-deprotecton" strategy for synthetic protecting group chemistry that links short and unreactive nucleic acid analogs to protecting groups which are then sequence-specifically deprotected with appropriate reagents linked to complementary nucleic acid analogs. In addition, we are interested in applying combinatorial approaches coupled with gene chip analysis (typically employed by the field of genomics) towards the comprehensive characterization and refinement of sequence-specific DNA binders[12] as well as towards the discovery of entirely new classes of DNA-binding molecules. Finally, we are interested in the design and implementation of triplet gene synthesis system that may allow the one-pot generation of DNA libraries encoding all possible alanine scanning, truncation, insertion, deletion, or recombined mutants of a protein or protein family of interest.